I thought that you use DataStreamer to load data into cache for some reason.
As I now see in gitter, you actually use ICacheStore.LoadCache.

So my next thought is that slowdown can be caused by GC because of low
memory.
How much memory one row takes? Can you show TransactionsDetail code, does
it have a lot of fields?
Can you attach to your process with jvisualvm.exe (part of Java JDK) and
see how much heap memory is actually used?
Another option is to look at Private Working Set in Task Manager.

Please also attach Ignite log.

On Wed, Mar 15, 2017 at 8:13 PM, Saifullah Zahid <saif.m...@gmail.com>
wrote:

> I tried that but it is throwing exception
>
> class org.apache.ignite.IgniteCheckedException: Cannot enable
> write-behind (writer or store is not provided) for cache:
>
> If I comment following settings
>
> ReadThrough = true,
> WriteThrough = true,
> WriteBehindEnabled = true,
>
> cache gets created but without loading data from database.
>
> thanks,
>
>
>
> On Wed, Mar 15, 2017 at 10:00 PM, Pavel Tupitsyn <ptupit...@apache.org>
> wrote:
>
>> Why use Spring?
>> I propose to do just one thing - comment out the CacheStoreFactory line.
>>
>> On Wed, Mar 15, 2017 at 7:30 PM, Saifullah Zahid <saif.m...@gmail.com>
>> wrote:
>>
>>> Hi,
>>>
>>> If I disable CacheStoreFactor then I have to provide the
>>> CacheConfiguration in spring xml
>>> But I don't know how to provide Database and table mapping in spring
>>> xml, I tried following configuration but it is failing.
>>> I am not sure if it is right way to do?
>>>
>>> <?xml version="1.0" encoding="UTF-8"?>
>>> <beans xmlns="http://www.springframework.org/schema/beans";
>>>        xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance";
>>>        xmlns:util="http://www.springframework.org/schema/util";
>>>        xsi:schemaLocation="http://www.springframework.org/schema/beans
>>>                            http://www.springframework.or
>>> g/schema/beans/spring-beans.xsd
>>>                            http://www.springframework.org/schema/util
>>>                            http://www.springframework.or
>>> g/schema/util/spring-util.xsd">
>>>
>>>     <bean id="ignite.cfg" class="org.apache.ignite.confi
>>> guration.IgniteConfiguration">
>>> <bean id="myDataSource" class="org.springframework.jdb
>>> c.datasource.DriverManagerDataSource">
>>> <property name="driverClassName" 
>>> value="com.microsoft.sqlserver.jdbc.SQLServerDriver"
>>> />
>>> <property name="url" value="jdbc:sqlserver://localh
>>> ost:1433;databaseName=TransactionsDB" />
>>> <property name="username" value="user1" />
>>> <property name="password" value="password123" />
>>> </bean>
>>>  <!-- Enabling ODBC. -->
>>>  <property name="odbcConfiguration">
>>> <bean class="org.apache.ignite.configuration.OdbcConfiguration"/>
>>>  </property>
>>> </bean>
>>> <property name="cacheConfiguration">
>>> <list>
>>> <bean class="org.apache.ignite.configuration.CacheConfiguration">
>>> <property name="dataSourceBean" value="myDataSource" />
>>> <property name="databaseSchema" value="dbo" />
>>>                     <property name="name" value="transactionsdetail"/>
>>>                     <property name="databaseTable"
>>> value="transactionsdetail" />
>>>             </bean>
>>> </list>
>>> </property>
>>> </beans>
>>>
>>> Getting following error
>>>
>>>  Failed to instantiate configuration from Spring XML
>>>
>>> Thanks,
>>> Saif
>>>
>>> On Wed, Mar 15, 2017 at 8:29 PM, Pavel Tupitsyn <ptupit...@apache.org>
>>> wrote:
>>>
>>>> Have you tried commenting out cache store as I descrived above?
>>>>
>>>> On Wed, Mar 15, 2017 at 6:10 PM, Saifullah Zahid <saif.m...@gmail.com>
>>>> wrote:
>>>>
>>>>> Hi,
>>>>>
>>>>> following is my load cache method
>>>>> if I comments out following line of code, it does not slows down
>>>>> act(GetKeyValue(item), item);
>>>>>
>>>>> public void LoadCache(Action<object, object> act, params object[] args)
>>>>> {
>>>>> using (DbConnection _connection = Create())
>>>>> {
>>>>> using (DbCommand command = _connection.CreateCommand())
>>>>> {
>>>>> MakeLoadCacheCommand(command, args);
>>>>> using (var reader = command.ExecuteReader())
>>>>> {
>>>>> while (reader.Read())
>>>>> {
>>>>> var item = new T();
>>>>> Map(reader, item); //populate object
>>>>> * act(GetKeyValue(item), item); //insert into cache*
>>>>> }
>>>>> }
>>>>> }
>>>>> }
>>>>> }
>>>>>
>>>>> Looks like issue is when inserting data into cache.
>>>>>
>>>>> Thanks,
>>>>> Saif
>>>>>
>>>>>
>>>>> On Wed, Mar 15, 2017 at 7:02 PM, Pavel Tupitsyn <ptupit...@apache.org>
>>>>> wrote:
>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> You have cache store configured, which probably causes the slowdown.
>>>>>> Please try to disable cache store (remove CacheStoreFactory from
>>>>>> config) and see if it makes any difference.
>>>>>>
>>>>>> On Wed, Mar 15, 2017 at 4:02 PM, Saifullah Zahid <saif.m...@gmail.com
>>>>>> > wrote:
>>>>>>
>>>>>>> Hi,
>>>>>>>
>>>>>>> I am facing an issue on cache loading, In the start cache loads
>>>>>>> quickly but after some time it becomes very slow almost 1 row per 
>>>>>>> second.
>>>>>>> There are about 4 million rows in a table.
>>>>>>> OS = Windows Server 2012
>>>>>>> RAM = 64 GB
>>>>>>> Node Heap config is 16 GB.
>>>>>>> Following is Cache configuration
>>>>>>>
>>>>>>> TransactionsDetailStore = m_cache.GetOrCreateCache<int,
>>>>>>> TransactionsDetail>(new CacheConfiguration("transactionssdetail",
>>>>>>> typeof(TransactionsDetail))
>>>>>>> {
>>>>>>> CacheStoreFactory = new TransactionsDetailStoreFactory("ApplicationDB",
>>>>>>> true),
>>>>>>> ReadThrough = true,
>>>>>>> WriteThrough = true,
>>>>>>> WriteBehindEnabled = true,
>>>>>>> KeepBinaryInStore = false,
>>>>>>> WriteBehindFlushThreadCount = 4,
>>>>>>> WriteBehindFlushFrequency = new TimeSpan(0, 0, 2),
>>>>>>> MemoryMode = CacheMemoryMode.OffheapTiered,
>>>>>>> OffHeapMaxMemory = 0,
>>>>>>> EvictionPolicy = new LruEvictionPolicy { MaxSize = 1000000 },
>>>>>>> WriteSynchronizationMode = CacheWriteSynchronizationMode.FullSync
>>>>>>> });
>>>>>>>
>>>>>>> Am I missing some configuration?
>>>>>>> Kindly see if anyone have idea?
>>>>>>>
>>>>>>> Thanks,
>>>>>>> Saif
>>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>
>

Reply via email to